Given that microscopic detail is necessary apparently to ID many snowfleas to species or even genus, chances are that many observations by nonexperts should probably be ratcheted back to a higher level, e.g. maybe to Collembola?

Okay, fine. Thanks.
This was in response to a posting in the forum some time ago in which users who noticed groups that were subject to overconfident IDs by the iNat photo recognition app were instructed to create a flag to explain the problem… only to be told later that identifiers should handle it themselves.